Expression of Secondary Metabolite Gene Clusters and Production of Secondary Metabolites in Three Nostoc strains Subjected to Deprivation on Nutrients and Competition

(1)

0 The Norwegian College of Fishery Science, UiT The Arctic University of Norway

Expression of Secondary Metabolite Gene Clusters and Production of Secondary Metabolites in Three Nostoc strains Subjected to Deprivation on Nutrients and Competition.

Oda Sofie Bye Wilhelmsen

Master thesis in Marine Biotechnology (May 2019) 60 credits

(2)

(3)

I

Acknowledgments

This master thesis was conducted at the Artic Marine Biology Institute in collaboration with Marbio, Norwegian College of Fisheries. The project lasted from January 2017 until June 2019 and concluded my master’s degree in Marine Biotechnology at UiT The Arctic University of Norway. I would like to thank my supervisors: Anton Liaimer and Espen Hansen. Special thanks to Anton Liaimer who has been there for all my hurdles along the way. Thanks to Rigmor Reiersen and Bente Lindgård for all your technical support. To my significant other, family and friends, thanks for all your love and support. And finally, thanks to all my fellow students for five great years.

Tromsø, May 2019

Oda Sofie Bye Wilhelmsen

(4)

II

Abstract

Cyanobacteria are a unique source of natural products, where most of them are synthesized by non-ribosomal peptide synthase, polyketide synthase, or a hybrid of these pathways. The identification of such biosynthetic genes responsible for the production of secondary metabolites is still a relatively unexplored area, and it remains many natural products for which a biosynthetic origin is unknown. Secondary metabolites from marine cyanobacteria have gained much attention the last decades, however few comprehensive studies on

secondary metabolites and their biosynthetic gene clusters from terrestrial cyanobacteria have been conducted.

Three terrestrial Nostoc spp. KVJ20, KVJ2, and KVJ10 were recently sequenced which allowed us to conduct genome-wide predictions by AntiSMASH of their potential to produce secondary metabolites. These strains were subjected to various cultivation conditions

including nutrient limitations and competition. Gene expression analysis by RT-qPCR of predicted gene clusters and the production of secondary metabolites by UPLC-HR-MS were conducted. Analysis of gene expression patterns revealed higher expression of several NRPS, PKS and RiPP genes in nutrient-deprived media, as well as confirming the present known function of the housekeeping genes; NifH, GvpC, and PilT. Most of the secondary metabolites found by UPLC-HR-MS were not identified, however a variety of Nostocyclopeptides,

Suomilide/Banyaside-like peptides, Anabaenopeptins, as well as Aeruginosin, Hapalosin, and Nosperin were recorded from extracts of the respective strains. The results from this thesis give valuable knowledge for further cultivation of terrestrial cyanobacteria, with the purpose of awaking cryptic gene clusters and identifying novel secondary metabolites. We have also suggested conditions most suitable for enrichment for identified compounds.

(5)

III

Table of content

ACKNOWLEDGMENTS ... I ABSTRACT ... II TABLE OF CONTENT ... III ABBREVIATIONS ... V

INTRODUCTION ... 1

A prokaryote with photosynthesis – The cyanobacteria ... 1

Morphology ... 2

Nutrition ... 3

Habitats ... 3

Nostocales ... 4

Production of secondary metabolites ... 5

Biosynthesis of secondary metabolites ... 6

Non-ribosomal peptide synthases ... 7

Hybrid pathways (NRPS/PKS) ... 8

Anabaenopeptins ... 8

Aeruginosins ... 9

Nostopeptolides & Nostocyclopeptides ... 11

Polyketide Synthases ... 12

Ribosomal peptide synthesis of complex peptides (RiPP) ... 12

Bioinformatics as a tool for finding novel nature products ... 15

Subject of study; Nostoc spp. KVJ20, KVJ10 & KVJ2 ... 16

AIM ... 17

METHODS AND EQUIPMENT ... 18

Workflow ... 18

Cultivation conditions. ... 20

Experimental setup ... 20

Bioinformatics ... 22

Primer design ... 22

Primer test ... 23

RNA extraction ... 24

cDNA synthesis ... 24

Reverse transcription quantitative PCR ... 25

Extraction of secondary metabolites ... 26

Medium extraction. ... 26

Cell extraction. ... 27

Mass Spectrometry ... 27

RESULTS ... 28

Homologous gene clusters ... 28

Nostoc KVJ20 ... 29

Nostoc KVJ2 ... 30

Nostoc KVJ10 ... 31

Observed changes associated with different cultivation conditions. ... 31

Allelopathic experiment ... 33

Gene expression patterns ... 34

Analysis of gene expression patterns in KVJ20 ... 35

Fold changes against respective controls for KVJ20 ... 35

(6)

IV

Fold changes against standard media (-N) for KVJ20 ... 36

Fold changes against respective controls for KVJ2 ... 38

Fold changes against standard media for KVJ2 ... 39

Fold changes against respective controls in KVJ10 ... 40

Fold changes against standard media (-N) for KVJ10 ... 42

Comparison of the gene expression patterns between homologous gene clusters ... 44

Gene expression patterns between homologous NRPS & PKS gene clusters ... 44

Gene expression patterns between homologous RiPP gene clusters ... 47

UPLC-HR-MS profiling and Metabolite identification ... 48

Chromatograms ... 48

TIC chromatogram KVJ20 ... 48

Summary of chemical profiles and the identified secondary metabolites ... 53

Relative peak intensity of identified secondary metabolites ... 56

Relative peak intensity in KVJ20 ... 56

DISCUSSION ... 60

Observed morphological changes and allelopathic interactions ... 60

Growth conditions affecting gene expression ... 62

General traits of housekeeping genes ... 62

Gene expression patterns in NRPS, PKS & RiPP gene clusters ... 63

Growth conditions altering the chemical diversity... 66

The chemical diversity in general traits. ... 66

Relative peak intensity and biotechnological potential. ... 68

Outlooks ... 71

CONCLUSIONS ... 73

REFERENCES ... 74

APPENDIX ... 82

Appendix 1: BG11 recipe... 82

Appendix 2: primers and locus. ... 83

Primers for KVJ20 ... 83

Appendix 3: Relative peak intensity in percentage. ... 86

Relative peak intensity in percentage for KVJ20 ... 86

Appendix 4: Fragmentation pattern examples. ... 89

(7)

V

Abbreviations

Aer Aeruginosin

Apt Anabaenopeptin

AvaK Akinete marker gene

cDNA Complementary DNA

dNTP Deoxyribonucleotide triphosphate ESI Electrospray ionization

FC Fold Change

GvpC Gas vesicle marker gene

Hap Hapalosin

Hgl Heterocyst glycolipids

KVJ10 Nostoc sp. KVJ10

Lan Lanthipeptide

m/z Mass-to-Charge

MDR Multidrug-Resistant

Mvd Microviridin

MQ Milli-Q Ultrapure Water

NCBI National Center for Biotechnology Information

Ncp Nostocyclopeptide

Ngn Nostoginin

NifH Nitrogen fixation marker gene NRPS Non-ribosomal peptide synthases

Nsp Nosperin

PilT Twitching motility marker gene PKS Polyketide Synthases

qTOF Quadrupole Time-of-Flight

RiPP Post ribosomally and post-translationally produced peptides

RT Retention time

RT-qPCR Reverse transcriptase-quantitative polymerase chain reaction

S/B Suomilide/Banyaside

Sid Siderophore

TIC Total Ion Chromatogram

UPLC-HR-MS Ultra-Performance Liquid Chromatography-High resolution-Mass spectrometry

Val-Val Valine-Valine bounds

n.d Not detected

(8)

1

Introduction

A prokaryote with photosynthesis – The cyanobacteria

Among the oldest organism on the earth, we find the cyanobacteria. Fossils of these unique microorganisms have been dated back to between 1,8 billion and 2,5 billion years ago

(Rasmussen et al., 2008; Schirrmeister et al., 2011), and today they are one of the largest and most significant group of bacteria inhabiting the earth.

The cyanobacteria have contributed to developing the earth as we know it, and their

production of oxygen helped oxygenate the earth about 2.5 billion years ago and contributed to making the earth inhabitable for bigger organisms like ourselves. They completely altered the course of evolution by the development of aerobic respiration, thus giving life to more complex lifeforms (Soo et al., 2017).

A well-known theory supposes that in the late Proterozoic era or the early Cambrian Period the cyanobacteria began to occupy some eukaryote cells, making nutrient supplies for the eukaryote host in exchange for shelter. This is known as the endosymbiosis theory and

supposes that the chloroplast in algae and plants originated from endosymbiotic cyanobacteria (Martin et al., 2015).

Cyanobacteria are often referred to as blue-green algae, and by eye resembles algae, but they are recognized as a major group of bacteria distinguished from other photosynthetic bacteria by their nature of their pigment system and by their performance of aerobic photosynthesis.

Even though they, in fact, are not algae, they show some similar features with algae and plants by having chlorophyll a and photosynthesis which leads to the development of oxygen

(Stanier et al., 1971).

Cyanobacteria consist of oxygenic photosynthetic prokaryote which has two photosystems (PSII and PSI), with H2O as photo reductant in the photosynthesis. All known cyanobacteria are photoautotrophic, which use CO2 as carbon source. The photopigments characteristic of cyanobacteria are chlorophyll a, phycobiliproteins (C-phycocyanin, allo-phycocyanin, and in some strains phycoerythrin), and in a variable array of carotenoids (O'Carra et al., 1980).

Where chlorophyll a and phycobiliproteins are their primary photosynthetic pigments (Waterbury, 2006)

(9)

2

Morphology

Throughout their long evolutionary history, the cyanobacteria have diversified into a variety of species with various morphologies, niche habitats and having unique interactions with other organisms (Foster et al., 2011; Freeman & Thacker, 2011). Cyanobacteria are developmentally and morphologically one of the most diverse groups of Prokaryotes, including unicellular, surface-attached, filamentous colony- and mat-forming species (Castenholz et al., 2001)

They range from simple unicellular forms, that reproduce by binary fission to complex filamentous forms that possess a variety of highly differentiated cell types, some forms are even capable of true branching (Gugger & Hoffmann, 2004; Schirrmeister et al., 2011). The recognition of cyanobacteria is mainly based on differences in the structure and development of five large sub-groups. These sub-groups do not, for the most part, correspond precisely to major taxa now recognized by phycologists; the sub-groups are defined as section I-V, section I-II include the unicellular cyanobacteria, and III-V include filamentous cyanobacteria

(Castenholz, 2015)

Members belonging to the section I form single cells and reproduce by binary fission or buddying and include members of the genus Synechocystis. Members belonging to section II either only reproduce by multiple fission, or by both binary fission and multiple fission, which gives rise to small daughter cells (baeocytes) and include members of the genus Dermocarpa and Xenococcus (Rippka et al., 1979).

While the filamentous cyanobacteria fill up section III, IV and section V, and they all consist of chains of cells called trichomes. In section III the trichrome is always composed of only vegetative cells, and division happens only on one plane for example, members of the genus Spirulina (Rippka et al., 1979). In section IV and V, the trichome will contain other

specialized cells, like the heterocyst in the absence of combined nitrogen. The members of section IV divide on only one plane and include members of the genus Nostoc, Anabaena and Cylindrospermum. The members of section V divide on more than one plane and include members of the genus’s Fischerella and Chlorogloeopsis (Rippka et al., 1979).

(10)

3

Nutrition

Cyanobacteria are photoautotrophic, using a photosynthetic apparatus that carries out

oxygenic plantlike photosynthesis with chlorophyll a as the primary photosynthetic pigment, and phycobiliproteins as auxiliary light-harvesting pigments (Waterbury, 2006).

Cyanobacteria are also capable of mixotrophy, a process in which a variety of organic

compounds, such as amino acids, that can not serve as sole carbon sources, are assimilated as a supplement to autotrophic CO2 fixation (Chojnacka & Noworyta, 2004; Mitra & Flynn, 2010). Many free-living cyanobacteria can fix dinitrogen, which permits them to exploit habitats low in combined nitrogen. The nitrogenase enzyme system responsible for the fixation of dinitrogen, is sensitive to oxygen, hence some of the filamentous cyanobacteria (IV-V) has developed highly differentiated cells, known as heterocysts where the fixation takes place. Since the nitrogenase enzyme is sensitive to oxygen, the heterocysts also lack the ability to carry out oxygenic photosynthesis (Waterbury, 2006).

Since the cyanobacteria are photoautotrophs, they can be grown in simple mineral media, supplemented with essential nutrients to support cell growth, such as sources of nitrogen, phosphorus and trace elements (Waterbury, 2006). Because of their minimum demand of nutrients, cyanobacteria can inhabit several diverse environments included extreme ones (Liengen & Olsen, 1997; Pushkareva et al., 2018).

Habitats

In addition to inhabiting a diverse range of aquatic, terrestrial and marine habitats,

cyanobacteria are also commonly found in more extreme environments; in Antarctica where they are found as cryptoendoliths in rocks (Blackhurst et al., 2005), in the dry desert (Potts &

Friedmann, 1981), in thermophilic lakes (Steunou et al., 2006), as well as unlikely habitats for phototrophs, such as Lava Caves (Saw et al., 2013). Most cyanobacteria are mesophilic and live in environments where the temperature range from freezing temperatures to 40°C.

However, their typical growth optimum is between 20 – 35°C and the maximum temperature for growth is typically below 45°C (Waterbury, 2006).

In addition to the diverse habitats occupied by free-living forms, both unicellular and filamentous cyanobacteria occur in symbioses with other prokaryotes, eukaryotic protist, metazoan and plants (Adams et al., 2013; Meeks, 1998). The cyanobacteria can inhabit different plant organs or tissues either intracellularly or externally (Santi et al., 2013). The ability of cyanobacteria to fix dinitrogen often plays an important role in these associations

(11)

4 (Meeks, 1998), where the plant acquires reduced nitrogen (N) from the cyanobacterial N2- fixation, whereas the cyanobacteria have shelter and are supplied with reduced carbon and other nutrients from the host (Rai et al., 2000). When nitrogen-fixation bacteria enter a symbiotic relation, both their morphology and their physiology changes dramatically. These changes involve primary a 5-10-fold increase in differentiation-rate of motile filaments called hormogonia, and of heterocysts in vegetative filaments that evolves from hormogonia

(Rodgers & Stewart, 1977).

Nostocales

Multicellular nitrogen-fixing cyanobacteria of the genus Nostoc are a common component of terrestrial microbial communities in a wide range of habitats, including subpolar and hot arid zones. Growth of Nostoc strains in both terrestrial and aquatic habitats often occur as

filaments inside a gelatinous matrix (Dodds et al., 1995).

Nostoc strains are nitrogen-fixating cyanobacteria belonging to the Nostocaceae family in the order Nostocales. Members of the order Nostocales belong to section IV, characterized as filamentous cyanobacteria with the ability to divide on one plane (Castenholz & Waterbury, 1989). Differentiation in Nostocales may result in the production of different types of

specialized cells; 1) heterocyst, 2) akinetes 3) hormogonia, specialized reproductive trichrome whose cells are morphologically different from vegetative cells, 4) tapered trichrome (Rippka et al., 1979; Waterbury, 2006). Nostoc is characterized by versatile physiology and displays one of the most complex life cycles among bacteria (Liaimer et al., 2016), as shown in Figure 1.

Figure 1: The lifecycle of Nostoc. Abbreviations: HC-Heterocyst, HIF-Hormogon Inducing Factors, HRF- Hormogon Repressing Factor. Illustration courtesy of Anton Liaimer.

(12)

5 Nitrogen fixation in these microorganisms is confined to specialized cells, heterocysts, which appear in semi-regular pattern between chains of vegetative cells, in response to a shortage of combined nitrogen in the environment. The photosynthesis in cyanobacteria takes place in the vegetative cells, where chlorophyll is stored in thylakoids in their cytoplasm (Heidrich et al., 2017). In addition, the life cycle of Nostoc includes the spore-like cells akinetes, which forms in response to nutrient deficiency other than Nitrogen, as well as motile multicellular

filaments hormogonia, the dispersal units (Rippka et al., 1979; Sarma et al., 2004).

Nitrogen-fixating cyanobacteria, primarily members of the Nostoc genus are often engaged in several symbiotic relationships with plants and fungi (Duggan & Adams, 2008; Meeks et al., 2002). Recent research has shown that which plant partner the cyanobacteria has in these symbiotic relations affects the production of secondary metabolites, doing so by down- regulation the biosynthesis of metabolites present in free-living stages, and inducing the production of other unknown products (Liaimer et al., 2015). Therefore the genus Nostoc, with its complex lifecycle and their diverse association with other organisms, is a unique model for the discovery of novel metabolites and compounds (Liaimer et al., 2016). In this paper, the focus will be directed towards members of Nostocales, and their production of secondary metabolites.

Production of secondary metabolites

Cyanobacteria produce an impressive array of natural products, with a vast diversity of structures (Burja et al., 2001; Harrigan & Goetz, 2002). This includes over 1100 secondary metabolites with exotic chemical structures reported from 39 genera of cyanobacteria, isolated from different geographic origins (Dittmann et al., 2015). Some of these secondary

metabolites are toxins that can cause severe health problems or even death for wild and domestic animals, or even for humans (de Figueiredo et al., 2006).

Cyanobacteria are without a doubt a rich source of natural products, represented as non- ribosomal peptides, polyketides, terpenes and alkaloids (Hoffmann et al., 2003; Pattanaik &

Lindberg, 2015; Taylor et al., 2014), as well as their production of fatty acids and variations of pigments (Los & Mironov, 2015; O'Carra et al., 1980).

For long, the research on natural products from cyanobacteria had its focus on toxins, and especially on the widespread hepatoxin microcystin 1 because of its threat to drinking-water (Dittmann & Wiegand, 2006). But from the early ’80s, several other interesting and promising compounds have been isolated from cyanobacteria, with diverse bioactivities and traits, for

(13)

6 example as antiproteases (Bister et al., 2004), anti-cancer (Costa et al., 2012), allelopathic (Liaimer et al., 2016), anti-viral (Kanekiyo et al., 2005), anti-inflammatory (Villa et al., 2010) and anti-infective traits (Balunas et al., 2010), as well as the production of neurotoxins (Choi et al., 2010). These findings are increasing the interest in the usage of cyanobacteria as a source of secondary metabolites that could be used not only as pharmaceuticals but also for fuel production, food, and other biotechnological applications, even the pigments may be utilized (Abed et al., 2009; Ducat et al., 2011).

In fact, there are few other groups of organisms that have shown such large and diverse chemical profile as cyanobacteria have, other than Myxobacteria and Streptomycetes.

Amazingly almost every new strain of cyanobacteria has its own unique set of secondary metabolites (Nunnery et al., 2010), which underlines that cyanobacteria are an excellent target for finding novel bioactive compounds.

Biosynthesis of secondary metabolites

A large part of the cyanobacterial secondary metabolites is peptides, or have peptide substances which are uniquely produced as a result of a naturally combined biosynthesis (Mandal & Rath, 2015; Nunnery et al., 2010; Ziegler et al., 1998). The productions of these peptides mainly happen by three different synthetic pathways; the non-ribosomal peptide synthases (NRPS), polyketide synthetases (PKS) and ribosomally where the peptides are post- translationally modified (RiPP) (Arnison et al., 2013; Koglin & Walsh, 2009).

Gene-clusters related to biosynthesis of these peptides have been assigned to an increasing number of natural products isolated from cyanobacteria (Jones et al., 2009; Welker & Von Döhren, 2006), where fascinating variations of enzymatic traits were observed, including many who rarely or never have been seen in other microorganisms (Kehr et al., 2011).

Most of the metabolites produced by cyanobacteria are assumed to be synthesized of non- ribosomal peptide synthase (NRPS), polyketide synthase (PKS)- or NRPS/PKS hybrid- pathways. This assumption is based on reported structures that are impossible to achieve by ordinary ribosomal synthesis (Welker & Von Döhren, 2006). Even though most of the metabolites from cyanobacteria are produced by these two pathways, several reports of biosynthetic pathways that start off with a ribosomally synthesized peptide that undergo posttranslational modification (RiPP) have been observed in cyanobacteria (Schmidt, 2010;

Schmidt et al., 2005). In this paper, the focus is directed towards these three pathways, and in the next sections, these pathways will be discussed in some detail.

(14)

7

Non-ribosomal peptide synthases

The non-ribosomal peptide synthase (NRPS) is a gigantic multi-domain enzyme which consists of different modules where each is responsible for the incorporation of an single amino acid. The basic module is built up of domains for adenylation (A), peptidyl carrier domain (PCP) and a condensation domain (C) (Koglin & Walsh, 2009; Kohli et al., 2001), where the amino acids are activated by domain A as adenylates, and the amino acids and the growing peptide chain are bound as thioesters to the pantetheine unit in the PCP domain, and finally peptide bond formation is catalyzed by the C domain (Conti et al., 1997; Lautru &

Challis, 2004).

The NRPs peptides are produced by enzyme complexes which are generated ribosomal by standard protein synthesis. The peptides produced by NRPS are often cyclic, have a high density of non-proteinogenic amino acids, and often contain amino acids connected by bonds other than peptide/disulfide bonds. The activation of amino acids in this multi-enzymatic process resembles the way amino acids are activated in ribosomal peptide synthesis, but the enzymes involved are neither structural or catalytical alike (Challis & Naismith, 2004).

There are also several alternative integrated modifying domains, among other the termination domain, where the release of the product with hydrolytic cleavage or intramolecular

cyclization is catalyzed by thioesterase (TE) activity and can be integrated in the C-terminus of the last module (Kohli et al., 2001; Lautru & Challis, 2004). Other alternative modifying domains include the epimerization domain, which epimerize aminoacyl and peptidyl

intermediates by thioester stage (Ansari et al., 2004). Other domains observed in NRPS with low frequencies include the oxidation domain, reduction domain and the formylation domain (Ansari et al., 2004). Integrated alternative associated enzymes are often a part of these modular complexes, such as dehydrogenases and halogenasaes (Wohlleben et al., 2009).

A distinct characteristic of the NRPS pathway is the ability to combine proteinogenic amino acids with non-proteinogenic amino acids, fatty acids, carbohydrates, and other building blocks to form complex molecules. The complex allows around 500 proteinogenic and non- proteinogenic amino acids (Walsh et al., 2013), in contrast to ribosomal peptide synthase which is limited to the 22 proteinogenic amino acids (Challis & Naismith, 2004).

The order of the NRPS modules corresponds with the order of amino acids in the final product, and the lack of specificity of the NRPS biosynthetic pathway contributes to the chemical diversity found in natural products from cyanobacteria (Fewer et al., 2007;

(15)

8 Stachelhaus et al., 1999). In fact, Nostocyclopeptide A1 is one of the few pure NRPS gene clusters that has been described in cyanobacteria (Becker et al., 2004).

Hybrid pathways (NRPS/PKS)

One main characteristic for the secondary metabolites found in cyanobacteria is the frequent mixture of NRPS and PKS modules in their biosynthetic pathways, and often within a single reading frame (Kehr et al., 2011; Welker & Von Döhren, 2006). Since there are many structural and catalytic similarities between these two pathways, it is only natural that one often finds hybrids of these. Both linear and cyclic peptides have been known to have hybrid pathways, and they have been found in both freshwater (Keishi Ishida et al., 2007; Tillett et al., 2000), marine (Chang et al., 2002; Zhang et al., 2017) and terrestrial cyanobacteria (Hoffmann et al., 2003; Magarvey et al., 2006).

The first NRPS/PKS hybrid pathway identified in cyanobacteria was the biosynthesis of the famous hepatoxin microcystin 1 from the freshwater cyanobacteria Microcystis aeruginosa (Nishizawa et al., 2000; Tillett et al., 2000). Other peptides with hybrid pathways include Cryptophycins which have been isolated from terrestrial strains of Nostoc either free-living or in symbioses with liches (Golakoti et al., 1995; Magarvey et al., 2006), and Barbamide and Jamaicamide, characterized from Lynbya majuscula (Chang et al., 2002; Edwards et al., 2004). In the following section, I am going to present some of the major peptide classes mainly from the genus Nostoc, produced by NRPS or hybrid NRPS/PKS pathways, where we have basic knowledge about the biosynthetic origin.

Anabaenopeptins

Anabaenopeptins are found across the phylum cyanobacteria and have among others been observed in the genera Nostoc, Anabaena, Planktothrix and Microcystis (Guljamow et al., 2017; Welker & Von Döhren, 2006). Anabaenopeptins are bioactive cyclic hexapeptides first isolated from the cyanobacteria Anabaena flosaquae (Harada et al., 1995), which are

characterized by a lysine in position 5 in the formation of a ring from a N-6-peptide bond between Lys and the carboxyl group of amino acids in position 6 (Harada et al., 1995).

Another trait of the Anabaenopeptins is that all the amino acids are in L-configuration, except the Lys in position 2 (Welker & Von Döhren, 2006). Anabaenopeptins vary in mass and have been observed with a mass range from 759 Da for Anabaenopeptin I (Murakami et al., 2000) to 956 Da for Oscillamide C (Sano et al., 2001).

(16)

9 Aeruginosins

Aeruginosins are a class of linear tetrapeptides commonly found in freshwater cyanobacteria and have an unusual hydroxy-phenyl lactic acid (Hpla) at the N-terminal and a 2-carboxy-6- hydroxyoctahydroindole (Choi) unit and an arginine derivate at the C terminus (Murakami et al., 1995). Most Aeruginosins are strong specific inhibitors of serine proteases (Ersmark et al., 2008; Keishi. Ishida et al., 1999). Aeruginosins has also been found in a wide range of Nostoc strains (Liaimer et al., 2016). Aeruginosin A is known to be produced by NRPS/PKS hybrid pathway (Keishi Ishida et al., 2007) presented in Figure 3, the two separate pathways NRPS and PKS and their typical domains are presented in Figure 2.

The loading module in the biosynthesis of Aeruginosin A has been predicted to activate phenylpyruvate which is reduced by an integrated KR domain to phenylacetate. The following adenylation domain activates the unusual Choi unit directly as a substrate (Keishi Ishida et al., 2007). However, the Aeruginosin enzyme complex doesn’t consist of a reductase or a

thioester domain, as shown in Figure 3, hence it is still unclear how the final product Aeruginosin A is released from the modular enzyme complex (Keishi Ishida et al., 2007).

(17)

10

Non-ribosomal peptide synthases (NRPS) Polyketide synthases (PKS)

Figure 2: The NRPS and PKS pathway and their typical domains. Abbreviations: C: condensation domain, A:

adenylation domain, PCP: peptidyl carrier protein, MT: Methyltransferase, E: epimerase, AT: Acyltransferase, ACP: Acyl carrier protein, KS: ketosynthase, KR: ketoreductase, DH: dehydrogenase, ER: enoyl reductase, TH:

thioesterase. PPT: 4’PPtase (PCP-specific/4’-phosphopantetheinyl cofactor transferase). Illustration modified from Jenke-Kodama et al. (2005); Kehr et al. (2011); Koglin and Walsh (2009)

Aeruginosin A, a known NRPS/PKS hybrid

Figure 3: Model for the biosynthesis of Aeruginosin A and predicted domains of AerA-I and supposed AerH, where we clearly see the hybrid nature of Aeruginosin A. Each square with rounded edges/circle represents an NRPS domain or tailoring function, and each square represents a PKS domain. Abbreviations: A, adenylation domain; KR, ketoreductase domain; ACP, acyl carrier protein; C, condensation domain; PCP; peptidyl carrier protein; E, epimerization domain. Illustration modified from Keishi Ishida et al. (2007).

(18)

11 Nostopeptolides & Nostocyclopeptides

Nostopeptolides (Nos) are branched acylated octapeptides with a heptapeptide lactone

structure (Figure 5) and are among others produced by the terrestrial cyanobacteria Nostoc sp.

GSV224 (Golakoti et al., 2000). The Nostopeptolide NRPS/PKS gene cluster was the first described for a terrestrial strain (Nostoc sp. GSV224), and has 3 NRPS genes (nosA, C and D) and one PKS gene (nosB) for acetate insertion and to genes (nosE and F) involved in

formation of 4-methyl proline (Hoffmann et al., 2003; Luesch et al., 2003), as well as one ABS transporter (nosG) (Hoffmann et al., 2003). Studies by Liaimer et al. (2015) have also shown that Nostopeptolides are involved in among regulation of hormogonia, and they serve a damper on hormogonia formation (Liaimer et al., 2016).

Nostocyclopeptides (Ncp) are cyclic heptapeptides (Figure 4), which possess a unique imino linkage in the macrocyclic ring (Golakoti et al., 2001). Nostocyclopeptides are structurally similar to Nostopeptolides, and the organization of gene clusters of Ncp and Nos shares similar traits, and there is a high degree of homology between the key genes in the operons (Becker et al., 2004). Studies have shown that multiple strains of Nostoc discovered both in soil and in symbiotic relations produce Nostocyclopeptide (Golakoti et al., 2001; Liaimer et al., 2016). Nostocyclopeptides may also be involved in regulation, similar to Nostopeptolides because of the high degree of homology (Liaimer et al., 2016).

Figure 4: Chemical structure of Nostocyclopeptide A1 and A2. The imine linkage in Nostocyclopeptide are highlighted in grey. Illustration courtesy of (Luesch et al., 2003)

Figure 5: Chemical structure of Nostopeptolide A1-A3. Illustration courtesy of (Luesch et al., 2003).

(19)

12

Polyketide Synthases

The bacterial polyketides are products of biosynthetic processes analogous to the fatty-acid biosynthesis. They are a remarkable class of natural products and in addition, to having an enormous range of functional and structural diversity, they have been proven to show antibiotic, anticancer, antifungal, antiparasitic and immunosuppressive properties (Singh et al., 2011; Staunton & Weissman, 2001). In contrast to the peptide-synthesizing enzymes, in PKS-systems different types of carboxylic acids are activated, assembled and then potentially modified (Kehr et al., 2011). All polyketide synthases use small acyl-coenzyme A (acyl CoA) units like propionyl, acetyl, malonyl or methyl malonyl CoA in sequential decarboxylative condensations reactions to form linear or cyclic carbon backbones (Staunton & Weissman, 2001).

The polyketide synthases have been classified by their structural resemblance with fatty acid synthases. Type I PKS are big modular proteins which carry all the active domains required for polyketide synthesis and are structurally like class I fungal and vertebrate fatty acids synthases. While in type II PKS the enzymes work repetitive and are typically involved in the biosynthetic of aromatic antibiotics in bacteria (Hopwood, 1997; Staunton & Weissman, 2001). The typical set of domains in an individual PKS consist of ketosynthase (KS), acyltransferase (A), and acyl carrier protein (ACP), in addition to the alternative domains ketoreductase (KR), dehydratase (DH), and enoyl reductase (ER) used to lead to another reduction state of keto-groups of polyketides (Hopwood, 1997; Staunton & Weissman, 2001), as seen in Figure 2. Curacin synthases are polyketides produced by strains of Lyngbya

majuscule, and Curacin A has shown potent cancer cell toxicity (Gerwick et al., 1994).

Ribosomal peptide synthesis of complex peptides (RiPP)

Even though the major share of cyanobacterial peptides has been proven to be synthesized non-ribosomally, complex and modified peptides can be synthesized independently from NRPS and PKS enzymes (Schmidt et al., 2005). The ribosomal precursor peptides are

typically built up of a leader peptide and a core peptide, which is transformed into the mature product (Figure 6) and associated post-translationally modifying enzymes (PTMs) catalyzes different types of macrocyclizations of the core peptide, and side-chain modification of amino acids (Oman & van der Donk, 2010).

In most RiPPs, the leader peptide or leader sequence is usually important for recognition by many of the post-translational modification enzymes and for export (Oman & van der Donk,

(20)

13 2010). Even though this biosynthesis of peptides is limited to the 22 proteinogenic amino acids (Challis & Naismith, 2004; Dittmann et al., 2015), many of these post-translational processing enzymes are highly tolerant to mutations in the core peptide which results in highly evolvable pathways (Arnison et al., 2013; Li et al., 2010; Oman & van der Donk, 2010). Hence, this group of peptides therefor still displays a high diversity and has a big bioactive potential, and cyanobacteria can be considered as one of the most prolific sources of ribosomal-produced natural products (Burja et al., 2001; Hetrick & van der Donk, 2017).

Figure 6: General biosynthetic pathway for RiPPs. Where the precursor peptide contains a core region that is transformed into the main product, with post-translational modifications guided by the leader peptide.

Illustration modified from Arnison et al. (2013).

Common features of RiPP biosynthesis include post-translational modification involving Cys residues, where sulfur chemistry converts the thiols of cysteines to disulfides (e.g.,

lanthipeptides & cyanobactins), thioethers (e.g., sactipeptides), thiazol(in)es (e.g.,

bottromycins), and finally sulfoxides (e.g., lanthipeptide & amatoxins) (Arnison et al., 2013).

Additionally, a common feature is the macrocyclization to increase metabolic stability and decrease conformational flexibility (Arnison et al., 2013), the typical macrocyclization of RiPPs is clearly shown in Figure 7-8.

The three biggest peptide families they produce are cyanobactins, lanthipeptides, and microviridins (Kehr et al., 2011). All known cyanobactins derive from cyanobacteria, and about 200 cyanobactins have been identified and evidence suggests that they might be present in 30% of cyanobacterial strains (Donia et al., 2008; Leikoski et al., 2010; Leikoski et al., 2009; Sivonen et al., 2010). The first ribosomal cyanobactin pathway which was discovered was the biosynthesis of Patellamides A and C, Figure 7, first isolated from Prochloron

(21)

14 didemni (Schmidt et al., 2005). Patellamides are pseudo-symmetrical cyclo-octapeptides which have proven to be moderately cytotoxic and have shown MDR reversing activity in cancer cells (Fu et al., 1998; Williams & Jacobs, 1993).

Figure 7: The chemical structure of Patellamide A and Patellamide C, illustration courtesy of Schmidt et al.

(2005)

The lanthipeptides are a big class of peptides which is produced by a variety of bacteria and is characterized by the presence of intramolecular lanthionine and methyl-lanthionine bridges formed by dehydration of serine/threonine, respectively, followed by intramolecular addition of cysteine thiols to the resulting dehydroamino acid (Cubillos-Ruiz et al., 2017).

Lanthipeptides have a variation of bioactivities, including antimicrobial ones commonly known as lantibiotics (Knerr & Donk, 2012).

The last big class of peptides is microviridins, which are the biggest known cyanobacterial oligopeptides. They are mainly found in the bloom of freshwater cyanobacteria (Fastner et al., 2001; Ziemert et al., 2008). The microviridins have an unusual cage-like structure and contain noncanonical lactone and lactam rings and several members of the family inhibit potent different serine-like proteases (Gatte-Picchi et al., 2014;

Ishitsuka et al., 1990). This unusual cage-like structure is shown in Microviridin G isolated from Nostoc Minutum

(Murakami et al., 1997), Figure 8. Figure 8: The chemical structure of Microviridin G, illustration courtesy of Murakami et al. (1997).

(22)

15

Bioinformatics as a tool for finding novel nature products

Bioinformatics has shown that the biosynthetic pathway of NRPS/PKS is an ancient part of (cyano)bacterial metabolism, thus cyanobacteria have been synthetizing secondary

metabolites long before higher animals and plant existed. Hence the intoxication of humans and other higher animals is most certainly not the selective force for the evolution of such a diverse repertoire of toxic or otherwise biologically active secondary metabolites from cyanobacteria (Berry et al., 2008). However, the production of these structures together with natural selection made it more advantageous to produce these unique structures (Welker &

Von Döhren, 2006).

The identification of biosynthetic genes responsible of the production of secondary metabolites is still a relatively unexplored area, and it remains many natural products for which a biosynthetic origin is unknown (Welker & Von Döhren, 2006). Hence knowledge about these pathways can reveal unique biochemical features, activities, and structures, as well as biotechnological applications (Dittmann et al., 2013; Hess, 2011).

Identifying putative biosynthetic gene clusters in the genome sequence protein in silico is generally a simple task with a BLAST search and further characterize and predict the

biosynthetic products by the enzymes encoded in the gene clusters with specialized software tools (Ingolfsson & Yona, 2008; Punta & Ofran, 2008; Weber, 2014). The most common software for predicting secondary metabolites and their corresponding gene clusters is the NRPS predictor (Röttig et al., 2011), BAGEL (van Heel et al., 2013) and AntiSMASH (Blin et al., 2015; Medema et al., 2011) where the latter combines several databases and software’s (BAGEL, NRPS predictor, NCBI BLAST+, Glimmer3, etc.) to identify and annotate

secondary metabolite biosynthesis gene-clusters in both bacterial and fungal genomes (Blin et al., 2015; Medema et al., 2011).

Even though all genomic studies like this will be limited to already known homologs, bioinformatics is still a good tool for finding novel natural products from cyanobacteria and guide us towards new scientific findings (Dittmann et al., 2013).

(23)

16

Subject of study; Nostoc spp. KVJ20, KVJ10 & KVJ2

The subject for this study is three cyanobacteria belonging to the genus Nostoc, the strains are further known as KVJ2, KVJ10 & KVJ20. Members of the genus Nostoc are included in section IV, filamentous cyanobacteria with division in only one plane. These sub-strains live in symbioses with the liverwort Blasia pusilla L and are found inside cavities in symbiotic organs called auricles in the outer cellular layer in the wort, where each cavity is infected by one single clone, visualized in Figure 9. The sub-strains were first collected at the plant school at Kvaløya (62° N 18,81°E) in Tromsø, in Northern Norway (Liaimer et al., 2016).

In a former study by Liaimer et al. (2016), it has been shown that KVJ2, KVJ10, and KVJ20 produced a spectrum of substances. In the same study, Banyaside/Suomilide like compounds were found in extracts of KVJ20, with m/z H⁺ on 899, 927, 997, 1045 and 1115, where the latter two where sulfated with a natural loss of 80 Da. Nostocyclopeptides with m/z H⁺757 were also reported in KVJ20 and KVJ2.

In the supernatant of extracts from KVJ10, Nosperin m/z 564 Da was identified, which is an unusual product of NRPS biosynthesis involving trans-acyltransferase. Aeruginosin with m/z H⁺ 889 was observed in KVJ2, and the fragmentation patterns of all these molecules shared a common feature of loss of either 176 or 162 Da, which is indicative of the presence of

glucuronic acid or a hexose, respectively (Kapuścik et al., 2013; Reinhold et al., 1995).

In the same study by Liaimer et al. (2016), it was shown preliminary results indicating cytotoxic properties in KVJ2 & KVJ20 against several cell lines. In another recent study by

Figure 9: The Liverwort Blasia pusilla L. in its natural habitat.

Strains of Nostoc are encircled in their auricles. Illustration courtesy of Liaimer et al. (2016).

(24)

17 Guljamow et al. (2017) three Anabaenopeptins were isolated from the KVJ2, with 828, 842 &

812 [M+H]⁺. Terrestrial cyanobacteria is as mentioned poorly explored as a source to novel natural products (Liaimer et al., 2016). Therefore, it is interesting to look further into the strains, as a potential source of novel secondary metabolites. Recently the draft genome of KVJ20, KVJ2 & KVJ10 has been released and deposited on NCBI with accession number:

NZ_LSSA00000000.1, NNBU00000000.1, and NNBT00000000.1 respectively.

Aim

I. Identify potential secondary metabolite gene clusters in Nostoc by using bioinformatic tools.

II. To design and test primers sets suitable for qPCR for the identified gene clusters.

III. Study the gene-expression in the predicted genes with help of reverse transcription qPCR in response to different nutrient conditions, as well an in interaction with other Nostoc strains.

IV. Establish a UPLC-HR-MS profile for KVJ2, KVJ10, and KVJ20 in different modes of cultivation.

V. Identifying potential changes in gene expression patterns as a response to different modes of cultivation.

VI. Identify connections between gene expression patterns and the chemical profile in the different strains in response to different modes of cultivation.

VII. Identify known peptides produced by the strains and suggest potential products with dereplication and examination of fragmentation patterns from the UPLC-HR-MS profile.

VIII. Identify which growth conditions would give a higher yield of interesting secondary metabolites and determine which mode of cultivation is best suited for

biotechnological applications.

(25)

18

Methods and Equipment

Workflow

In this study, we wanted to analyze gene expression and production of secondary metabolites, with the help of bioinformatic tools as AntiSMASH. The workflow started with predicting secondary metabolites in the strains, by depositing the genomes on AntiSMASH, before making primers for selected gene clusters. The exponentially growing cyanobacterial

suspension was then subjected to nutrient deprivations, as well as subjected to competition to other Nostoc strain before extraction of secondary metabolites and gene expression analysis was conducted. In Figure 10 a shortened schematic representation of the workflow is shown; a more detailed description is described in the following chapter.

The biosynthesis of cyanobacterial secondary metabolites consumes a great deal of metabolic energy, hence clues to the function of these compounds may be revealed by exploring and altering the growth conditions under which they are produced (Briand et al., 2016). There is considerable evidence that some cryptic or poorly expressed secondary metabolites can be more activated under stress responses in bacteria (Yoon & Nodwell, 2014), which is the underlining basis for this project. By exposing the bacteria to different nutrient conditions, we hoped this would interfere with both the gene expression and the production of secondary metabolites.

To determine the gene expression of selected secondary metabolite genes chosen from bioinformatic analysis, we chose to use reverse transcriptase-quantitative PCR (RT-qPCR), which uses the same basic principles as the basic PCR technique. But instead of having amplicons after the PCR cycle as an endpoint in qPCR, the amount of DNA is measured after each cycle by using fluorescent markers that are incorporated into the PCR product (Heid et al., 1996; Martins & Vasconcelos, 2011). Thus, the increase in fluorescent signal is directly proportional to the number of amplicons generated (Heid et al., 1996). By using RT-qPCR the RNA transcript can be measured by converting the RNA template to its complementary DNA (cDNA) using reverse transcriptase, before being used as a template in the PCR (Bustin, 2002). Hence, we can measure the amount of RNA transcript in real time and quantify gene expression. So far, no published research on the relative expression of the complete set of NRPS, PKS and RiPP both with unknown and known products have been conducted on KVJ2, KVJ10, and KVJ20.

(26)

19 Ultra-high pressure liquid chromatography (UPLC) in combination with time-of-flight (TOF) mass spectrometry detectors have proven to be a very efficient tool for both the identification of new natural products and the dereplication of natural product extracts (Grata et al., 2008;

Jean-Luc et al., 2010). By using this sophisticated analytic method, we could easily identify and assess compounds present in the different extracts, as well as compare the chemical profile of the different compounds.

The complete “pipeline” was executed for KVJ20 by me, as well as extraction of secondary metabolites from KVJ2 and mass spectrometry analysis. Gene expression data from KVJ2 and KVJ10 were obtained from earlier Bachelor students, with some additional qPCR runs for quality checking. Mass spectrometry data for KVJ10 was provided by my supervisor. All extracts and data have been treated with the same conditions and the same methods for extractions.

Figure 10: Workflow in general traits. MeOH: methanol RT-qPCR: Reverse transcriptase- quantitative Polymerase chain reaction. UPLC-HR-MS: Ultra-Performance Liquid Chromatography-High resolution-Mass spectrometry.

(27)

20

Cultivation conditions.

For cultivation of KVJ20, 2 L of BG110 and BG11 medium were made by recipe from stock solution, first described in Stanier et al. (1971). The BG110 medium is a well-used medium, which lacks sodium nitrate and is used for cultivation of cyanobacteria capable of dinitrogen fixation, while BG11 has sodium nitrate added. All the solutions being used for cultivation were autoclaved, to maintain a sterile environment for growth.

Liquid media was made by adding the stock solution in order from I-IV (BG110)/VII (BG11) in a 1000-fold dilution in 2 L of Milli-Q water in an Erlenmeyer flask. The stock solutions we used are listed in Appendix 1: BG11 recipe. Solid plates were made from the same stock recipe with 1% added agar and poured on plates.

For cultivation 50 mL BG110/BG11liquid medium was added to six 100 mL sterile

Erlenmeyer flask, with a scoop of bacteria from the sub-strain KVJ20-G re-isolated as pure culture from symbiosis with Gunnera manicata. This sub-strain still contains most of the same properties as the wildtype, in contrast to sub-strains maintained in liquid cultures. The flasks were incubated for 1-3 weeks at standard conditions at 23°C with constant light 30 μmol m⁻²s⁻¹ (36W/77 Osram Fluora) and shaking.

Experimental setup

After cultivation the six exponentially growing (2 weeks old) cultures were aliquoted to 50 mL Falcon tubes, and the bacteria cells were collected by centrifugation (Eppendorf

Centrifuge 5804 R) at 5000 rpm for 5 minutes, and then followed by two times wash in fresh BG110/BG11medium. All pellets were combined, followed by an additional wash. The cell material was then homogenized in a small amount by a 0.4 mm syringe (BDMicrolance™).

Solid medium was prepared both with and without sodium nitrate from stock solution as previously described. Each agar plate was spotted with ten 20 µL drops of the cyanobacterial suspension, which later gave rise to colonies. Allelopathy experiments were set up by adding five drops of one Nostoc strain on one side of the agar plate, and five drops of another Nostoc strain on the other side, as shown in Figure 11. Each agar plate was covered in parafilm and cultivated at standard

conditions until visible results.

Figure 11: Illustration for how allelopathy experiments were conducted.

(28)

21 The remainder suspension was redistributed to the original cell density in 20 mL of the liquid media with and without nitrogen and in addition in phosphor-deficient and iron-deficient medium. Phosphor-deficient and iron-deficient media were simply made by not adding stock solution I and IV respectfully to the media. The starvation cultures were incubated under standard condition for 2 weeks. Agar plates and liquid media cultivations were made in triplicates, and all work was performed on a sterile bench to prevent contamination. The experimental set up is presented in Figure 12.

After one week of growth, samples from all cultures were taken for microscopy to study morphology and possible morphological differences in bacterial development and growth.

Microscopy was done with Leica DFC420 fluorescence microscope and samples were looked at under brightfield with a 40x objective.

The cultures from liquid media, were transferred to 50 mL Falcon tubes and centrifuged for 5 minutes at 5000 rpm, and the supernatant was collected and stored at - 20°C for later

extraction. Each pellet from the liquid media was divided into two parts, one for later secondary metabolite extraction and the other for RNA extraction, preserved in 300 μL of RNAlater™ Stabilizer Solution (ThermoFisher), and stored at - 20°C. The colonies from agar plates were collected by inoculation loop and divided into two equal parts as described above, one part for RNA extraction and one for secondary metabolite extraction.

Figure 12: The experimental setup of the cultivation of the cyanobacteria.

(29)

22

Bioinformatics

The genome sequences for KVJ20, KVJ2 and KVJ10 were retrieved from NCBI, with accession number: NZ_LSSA00000000, NNBU00000000.1, and NNBT00000000.1 respectively, and submitted to the specialized database AntiSMASH, which perform

antibiotics and secondary metabolite analysis, and allows the analysis of secondary metabolite biosynthesis gene clusters (Blin et al., 2015; Medema et al., 2011). In addition, a search in the Norine prediction tool (Pupin et al., 2015) was conducted for KVJ10. From the produced result from the database search, the different gene clusters were analyzed and one open

reading frame from each cluster was chosen for primer design. Additionally, BLAST analyses were run with Nostoc punctiforme ATCC29133 as query in order to find genes for

Nitrogenase (NifH), akinetes marker gene (AvaK), gas vesicle protein C (GvpC) and Ribonuclease P RNA gene (RnpB), motility protein (PilT). No GvpC gene was found in KVJ2, hence PilT was used instead.

Primer design

The genes that were chosen for primer design for the NRPS clusters was an NRPS gene, for PKS cluster a PKS gene, for RiPP clusters an ABC transporter and for the microviridins, a ribosomal modification protein (RimK) was chosen. The sequences of the target genes were exported to a target-specific primer designing software tool called Primer-BLAST (Ye et al., 2012). Primer design settings were set accordingly to suggestions by BioRad for Real-Time PCR, namely:

˃ Primer length: 20bp

˃ Amplicon size: 150-250bp

˃ GC content: ~50%

˃ Optimal Tm: 60°C

˃ Avoid repeats (e.g. ATATATAT) and long runs (e.g. CCCC) they may cause mispriming.

˃ Verify specificity using tools such as the Basic Local Alignment Search Tool

The primers designed for KVJ20, KVJ2, and KVJ10 are listed in Table 26-28 in Appendix along with gene clusters, possible product, locus and accession number from NCBI.

(30)

23

Primer test

All primers were tested before being used for real-time PCR using Whatman™ FTA® card technology, and were stored at -20°C. The primers were diluted to a concentration of 100 µM in sterile double-distilled water (MilliQ water). The suspension of Nostoc sp. KVJ20 was pipetted onto a CloneSaver^TM FTA® Card (Whatman™) and treated according to the instructions.

Paper circles containing cyanobacteria were pushed out from the card and into small PCR Eppendorf tubes, before being cleaned with FTA cleaning solution, two times (à five minutes) and then with TE^-1 two times (à five min). The Mastermix for PCR was made from recipe shown in Table 1, and the exact amounts of components were calculated up to a number of reactions that were run each time. All components in the Mastermix were stored at -20°C.

Before running PCR 25 µL of Mastermix added to each Eppendorf tube containing paper circles with bacteria. The PCR program that was used in primer testing is listed in Table 2.

After completing the PCR, a 1% agarose gel was run with 90 Volts in order to test if products were made. For visualization under UV-light ethidium bromide (EtBr) was added to the gel, and 1x TAE buffer was used. To the PCR tubes, 5 µL of 6x Loading DNA dye (Fermentas) was added. The results were analyzed by ChemiDoc^TMMP Imaging System.

Table 1: Mastermix x1 for PCR (25 µL) reaction Table 2: Polymerase chain reaction cycles.

(31)

24

RNA extraction

All samples were first centrifuged at 25000 rpm for 5 min to discard RNAlater™ stabilizer solution. The RNA extraction took place under a fume hood. The samples were homogenized before 1 mL of TRI Reagent® (Sigma-Aldrich) was added to each Eppendorf tube, and incubated at 95°C for 5 minutes, followed by 5 minutes on ice, before 100 µL of bromochloropropane was added to each tube and mixed vigorously. The samples were incubated for 10 minutes in room temperature with bromochloropropane, before being centrifuged at 12000g at 4°C for 5 minutes. After this centrifugation, the samples had

separated into three different layers. The upper layer was transferred to a new Eppendorf tube and an equal volume of isopropanol was added, and the samples were incubated at room temperature for 15 minutes and again centrifuged at 12000 g at 4°C for 10 minutes.

To wash the RNA the supernatant was discarded and 1 mL 75% Ethanol (EtOH) was added, followed by centrifugation at 8000g at 4°C for 5 minutes. The supernatant was removed, and the pellet was air-dried and dissolved in 50 µL of DEPC treated water supplemented by 2.5 units of SUPERrase• In™ RNase Inhibitor. The samples were treated with DNAse, by adding 5.3 µL 10x DNAse buffer (0.1 x sample volume) and 1 µL DNAse, followed by incubation for 25 minutes at 37°C. After incubation 6 µL of inactivation reagent was added

(approximately 0.1x sample volume) and mixed, before incubation for 2 minutes at room temperature and centrifuged at 10000g for 2 minutes. After centrifugation, the supernatant was transferred to a new tube, and concentration and purity of RNA were measured using NanoDrop™ 2000 Spectrophotometer. A 1% agarose gel (as described earlier) was run with 5 µL RNA to verify that the RNA was intact and that there was no DNA contamination in the samples.

cDNA synthesis

The SuperScript™ II Reverse Transcriptase (RT) was used to reverse transcribe the RNA to cDNA. In the first step, four compounds were added to PCR tubes: 1 µL Hexamer primer, 4 µL dNTP, 1 µg RNA and the amount of DEPC H2O was adjusted to fit the maximum volume of 12.5 µL after RNA was added. The samples were incubated for 5 minutes at 65°C on a heat block, followed by a quick cool down to 4°C on ice and spun down with a centrifuge to collect condensation, before 4 µL of 5x Second Strand Buffer (Invitrogen™) and 2 µL 0.1 M USB Dithiothreitol (DTT) (Thermo Scientific™) was added.

(32)

25 The samples were then gently mixed and centrifuged, before another incubation at 25°C for 2 min. Then 1 µL SuperScript II™ RT and 1 µL dH2O was added and the tubes were incubated for 10 minutes at 25°C, then for 50 min at 42°C followed by an inactivation period for 15 minutes at 70°C. The samples were stored at -80°C until use.

Reverse transcription quantitative PCR

We chose to run two technical duplicates for each biological sample, and each growth condition was represented by two biological replicates (two best RNA samples for each condition). All primer pairs were tested on each sample, and the RnpB primers were used as an internal housekeeping reference. Reactions x2 were set up using the SsoFast™ EvaGreen®

Supermix and Mastermix was made for each cDNA sample tested, Table 3. All the cDNA samples were diluted x100 in order to get appropriate cDNA concentration for the reaction.

The primers were also diluted x40.

The reactions were set up in 96 well plates. The CFX96^TM Real-Time PCR Detection System (Bio-Rad) was used to perform qPCR reactions. Data were analyzed with the CFX

Manager^TM Software program (Bio-Rad). The qPCR program is presented in Table 4.

The same threshold line was used on all runs, and relative expression was calculated by the formula: RE (relative expression) = 2-(cQgene – cQrnpB)

, with standard deviation (STDV), and fold changes in expression by the formula: FC (fold changes) = 2^-(∆cQexp - ∆cQref)

.

For graphic presentation, the fold changes below one were converted to negative values by -1/FC.

Table 3: Mastermix for qPCR reactions.

Mastermix x 1 reaction

SsoFast™ EvaGreen® Supermix 10 µL

2.5 µM reverse and forward primers 4 µL

cDNA 6 µL

Table 4: qPCR cycles.

Stage Temp Time Cycles

First denaturation 95°C 30 sec

Denaturation 95°C 5 sec X40

Annealing/extension 60°C 5 sec

Melt curve 65°C- 95°C 5 sec every 0,5°C

(33)

26

Extraction of secondary metabolites

Medium extraction.

The nonionic resin Amberlite® XAD7HP (Sigma Aldrich) was used to extract the content from the supernatant after centrifugation to get rid of cell residue. 60 g of Amberlite®

XAD7HP was conditioned by rinsing it with 2L dH2O three times, just enough100% MeOH to cover the resin grains once and left for 30 minutes, and finally washed with 2 L dH2O another three times. Once the Amberlite® XAD7HP was thoroughly cleaned it was air-dried covered with aluminum foil to avoid dust, after drying the Amberlite® XAD7HP was distributed to six 250 mL Erlenmeyer flasks.

The supernatant was centrifuged to remove any cell residue, and then added to the flasks containing 60g dry Amberlite® XAD7HP, the flask opening was covered with aluminum foil, the supernatant-resin mixture was stirred at low pace (150 rpm) overnight. Compounds

present in the medium should have adhered to the Amberlite® XAD7HP grains.

The supernatant-resin mixture was filtered in small amounts through a porcelain filter with a Whatman® quality 1 filter paper on top (90 mm Ø x 100), and the Amberlite® XAD7HP was collected and added to a 250 mL Erlenmeyer flask containing 50 mL of 100% MeOH and stirred at slow pace overnight. MeOH replaced the compounds attached to the Amberlite®

XAD7HP grains and washed out the organic compounds.

A rotary evaporation device (Laborota 4011, Heidolph™ rotavapor system) was used to dry the solution. The device was set on auto, with approximately < 37°C and 150 rpm. When working on the rotavapor system it is important to ensure that the solution doesn’t boil. The distilled MeOH was disposed of in appropriate waste bottles and left under the fume hood for evaporation.

The dry matter was first dissolved in 3.3 mL of MilliQ, then 3.3 mL 50% MeOH followed by 3.3 mL 100% MeOH, and the fractions were combined. The combined solution was aliquoted into 10 mL glass tubes and evaporated to dryness in SpeedVac Plus SC210A (Savant™) coupled with Refrigerated Condensation Trap RT400 (Savant™).

When dry the solids were washed out with 250 µL MQ, 250 µL 50% MeOH and then 250 µL 100% MeOH, before being collected in HPLC Vials. The total dissolved matter is 750 µL in 1:1 MeOH:MQ.

(34)

27

Cell extraction

The cells were harvested from 2 weeks old growing cultures (BG11/-Fe/-P, BG110/-Fe/-P), and the pellet was transferred to 15 mL falcon tubes. 5 mL 100% MeOH was added to the pellet before the samples were sonicated in a Branson Sonifier SFX250® to homogenize the samples, to release any compounds retained inside of the cells. Parameters of the sonicator were set to; 3 min, duty cycle 30 and output control to 5. The cycles were repeated when necessary. After sonication, the samples were centrifuged at 1500 rpm for 5 minutes, and the supernatant was collected. Then 3 mL of 50% MeOH was added to the pellet and centrifuged again, and the supernatant was collected. Finally, 3 mL of MQ water was added and the sample was centrifuged, and the supernatant was collected. The fractions were combined.

Then the combined fractions were filtered with Acrodisc® Syringe Filters 0.45 µm, 25 mm before transferring the fractions to 10 mL glass tubes. The fractions were evaporated to dryness in SpeedVac Plus SC210A (Savant™) coupled with Refrigerated Condensation Trap RT400 (Savant™). The dry solids were washed as described for medium extraction.

Extraction of colonies from solid media were performed as described above.

Mass Spectrometry

15 µL from the different fractions were placed in two UPLC tubes with 100 µL 50% MeOH to dilute the samples. UPLC-HR-MS analysis was performed on the samples using a Waters Acquity I-class UPLC system (Milford, MA, USA) interfaced with a PDA Detector and a VION IMS-qTOF, using electrospray ionization (ESI) in positive mode, wavelengths from 190-500 nanometers were detected. VION IMS- qTOF conditions for UPLC-HR-MS analysis includes capillary voltage (0.80kV), cone gas (50 L/h), desolvation temperature (350°C), desolvation gas (800 L/h), source temperature (120°C) and acquisition range (m/z 50–2000) The system was controlled and data was processed using UNIFI 1.8.2 (Waters).

Chromatographic separation was performed with an BEH C18 1.7 µm (2.1 × 100 mm) column (Waters) maintained at 40°C. Selected peaks were dereplicated using MarinLit, ChemSpider, and Dictionary of Natural products, as well as extensive literature searches.